2 Unix SMB/CIFS implementation.
3 Samba utility functions
5 Copyright (C) Andrew Tridgell 1992-2001
6 Copyright (C) Simo Sorce 2001-2002
7 Copyright (C) Martin Pool 2003
8 Copyright (C) James Peach 2005
10 This program is free software; you can redistribute it and/or modify
11 it under the terms of the GNU General Public License as published by
12 the Free Software Foundation; either version 2 of the License, or
13 (at your option) any later version.
15 This program is distributed in the hope that it will be useful,
16 but WITHOUT ANY WARRANTY; without even the implied warranty of
17 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 GNU General Public License for more details.
20 You should have received a copy of the GNU General Public License
21 along with this program; if not, write to the Free Software
22 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
26 #include "system/iconv.h"
29 #include "lib/ldb/include/ldb.h"
33 * @brief String utilities.
37 * Get the next token from a string, return False if none found.
38 * Handles double-quotes.
40 * Based on a routine by GJC@VILLAGE.COM.
41 * Extensively modified by Andrew.Tridgell@anu.edu.au
43 _PUBLIC_ BOOL next_token(const char **ptr,char *buff, const char *sep, size_t bufsize)
54 /* default to simple separators */
58 /* find the first non sep char */
59 while (*s && strchr_m(sep,*s))
66 /* copy over the token */
67 for (quoted = False; len < bufsize && *s && (quoted || !strchr_m(sep,*s)); s++) {
76 *ptr = (*s) ? s+1 : s;
83 Case insensitive string compararison
85 _PUBLIC_ int strcasecmp_m(const char *s1, const char *s2)
87 codepoint_t c1=0, c2=0;
91 c1 = next_codepoint(s1, &size1);
92 c2 = next_codepoint(s2, &size2);
101 if (c1 == INVALID_CODEPOINT ||
102 c2 == INVALID_CODEPOINT) {
103 /* what else can we do?? */
104 return strcasecmp(s1, s2);
107 if (toupper_w(c1) != toupper_w(c2)) {
116 Case insensitive string compararison, length limited
118 _PUBLIC_ int strncasecmp_m(const char *s1, const char *s2, size_t n)
120 codepoint_t c1=0, c2=0;
123 while (*s1 && *s2 && n) {
126 c1 = next_codepoint(s1, &size1);
127 c2 = next_codepoint(s2, &size2);
136 if (c1 == INVALID_CODEPOINT ||
137 c2 == INVALID_CODEPOINT) {
138 /* what else can we do?? */
139 return strcasecmp(s1, s2);
142 if (toupper_w(c1) != toupper_w(c2)) {
157 * @note The comparison is case-insensitive.
159 _PUBLIC_ BOOL strequal(const char *s1, const char *s2)
166 return strcasecmp_m(s1,s2) == 0;
170 Compare 2 strings (case sensitive).
172 _PUBLIC_ BOOL strcsequal(const char *s1,const char *s2)
179 return strcmp(s1,s2) == 0;
184 Do a case-insensitive, whitespace-ignoring string compare.
186 _PUBLIC_ int strwicmp(const char *psz1, const char *psz2)
188 /* if BOTH strings are NULL, return TRUE, if ONE is NULL return */
189 /* appropriate value. */
192 else if (psz1 == NULL)
194 else if (psz2 == NULL)
197 /* sync the strings on first non-whitespace */
199 while (isspace((int)*psz1))
201 while (isspace((int)*psz2))
203 if (toupper((unsigned char)*psz1) != toupper((unsigned char)*psz2)
210 return (*psz1 - *psz2);
215 NOTE: oldc and newc must be 7 bit characters
217 _PUBLIC_ void string_replace(char *s, char oldc, char newc)
221 codepoint_t c = next_codepoint(s, &size);
230 Trim the specified elements off the front and back of a string.
232 _PUBLIC_ BOOL trim_string(char *s,const char *front,const char *back)
239 /* Ignore null or empty strings. */
240 if (!s || (s[0] == '\0'))
243 front_len = front? strlen(front) : 0;
244 back_len = back? strlen(back) : 0;
249 while (len && strncmp(s, front, front_len)==0) {
250 /* Must use memmove here as src & dest can
251 * easily overlap. Found by valgrind. JRA. */
252 memmove(s, s+front_len, (len-front_len)+1);
259 while ((len >= back_len) && strncmp(s+len-back_len,back,back_len)==0) {
260 s[len-back_len]='\0';
269 Find the number of 'c' chars in a string
271 _PUBLIC_ size_t count_chars(const char *s, char c)
277 codepoint_t c2 = next_codepoint(s, &size);
278 if (c2 == c) count++;
286 Safe string copy into a known length string. maxlength does not
287 include the terminating zero.
289 _PUBLIC_ char *safe_strcpy(char *dest,const char *src, size_t maxlength)
294 DEBUG(0,("ERROR: NULL dest in safe_strcpy\n"));
299 /* We intentionally write out at the extremity of the destination
300 * string. If the destination is too short (e.g. pstrcpy into mallocd
301 * or fstring) then this should cause an error under a memory
303 dest[maxlength] = '\0';
304 if (PTR_DIFF(&len, dest) > 0) { /* check if destination is on the stack, ok if so */
305 log_suspicious_usage("safe_strcpy", src);
316 if (len > maxlength) {
317 DEBUG(0,("ERROR: string overflow by %u (%u - %u) in safe_strcpy [%.50s]\n",
318 (uint_t)(len-maxlength), (unsigned)len, (unsigned)maxlength, src));
322 memmove(dest, src, len);
328 Safe string cat into a string. maxlength does not
329 include the terminating zero.
331 _PUBLIC_ char *safe_strcat(char *dest, const char *src, size_t maxlength)
333 size_t src_len, dest_len;
336 DEBUG(0,("ERROR: NULL dest in safe_strcat\n"));
344 if (PTR_DIFF(&src_len, dest) > 0) { /* check if destination is on the stack, ok if so */
345 log_suspicious_usage("safe_strcat", src);
348 src_len = strlen(src);
349 dest_len = strlen(dest);
351 if (src_len + dest_len > maxlength) {
352 DEBUG(0,("ERROR: string overflow by %d in safe_strcat [%.50s]\n",
353 (int)(src_len + dest_len - maxlength), src));
354 if (maxlength > dest_len) {
355 memcpy(&dest[dest_len], src, maxlength - dest_len);
361 memcpy(&dest[dest_len], src, src_len);
362 dest[dest_len + src_len] = 0;
367 Paranoid strcpy into a buffer of given length (includes terminating
368 zero. Strips out all but 'a-Z0-9' and the character in other_safe_chars
369 and replaces with '_'. Deliberately does *NOT* check for multibyte
370 characters. Don't change it !
373 _PUBLIC_ char *alpha_strcpy(char *dest, const char *src, const char *other_safe_chars, size_t maxlength)
377 if (maxlength == 0) {
378 /* can't fit any bytes at all! */
383 DEBUG(0,("ERROR: NULL dest in alpha_strcpy\n"));
393 if (len >= maxlength)
396 if (!other_safe_chars)
397 other_safe_chars = "";
399 for(i = 0; i < len; i++) {
400 int val = (src[i] & 0xff);
401 if (isupper(val) || islower(val) || isdigit(val) || strchr_m(other_safe_chars, val))
413 Like strncpy but always null terminates. Make sure there is room!
414 The variable n should always be one less than the available size.
417 _PUBLIC_ char *StrnCpy(char *dest,const char *src,size_t n)
426 while (n-- && (*d++ = *src++))
434 Routine to get hex characters and turn them into a 16 byte array.
435 the array can be variable length, and any non-hex-numeric
436 characters are skipped. "0xnn" or "0Xnn" is specially catered
439 valid examples: "0A5D15"; "0x15, 0x49, 0xa2"; "59\ta9\te3\n"
443 _PUBLIC_ size_t strhex_to_str(char *p, size_t len, const char *strhex)
446 size_t num_chars = 0;
447 uint8_t lonybble, hinybble;
448 const char *hexchars = "0123456789ABCDEF";
449 char *p1 = NULL, *p2 = NULL;
451 for (i = 0; i < len && strhex[i] != 0; i++) {
452 if (strncasecmp(hexchars, "0x", 2) == 0) {
453 i++; /* skip two chars */
457 if (!(p1 = strchr_m(hexchars, toupper((unsigned char)strhex[i]))))
460 i++; /* next hex digit */
462 if (!(p2 = strchr_m(hexchars, toupper((unsigned char)strhex[i]))))
465 /* get the two nybbles */
466 hinybble = PTR_DIFF(p1, hexchars);
467 lonybble = PTR_DIFF(p2, hexchars);
469 p[num_chars] = (hinybble << 4) | lonybble;
479 * Parse a hex string and return a data blob.
481 _PUBLIC_ DATA_BLOB strhex_to_data_blob(const char *strhex)
483 DATA_BLOB ret_blob = data_blob(NULL, strlen(strhex)/2+1);
485 ret_blob.length = strhex_to_str((char *)ret_blob.data,
494 * Routine to print a buffer as HEX digits, into an allocated string.
496 _PUBLIC_ void hex_encode(const unsigned char *buff_in, size_t len, char **out_hex_buffer)
501 *out_hex_buffer = smb_xmalloc((len*2)+1);
502 hex_buffer = *out_hex_buffer;
504 for (i = 0; i < len; i++)
505 slprintf(&hex_buffer[i*2], 3, "%02X", buff_in[i]);
509 Check if a string is part of a list.
511 _PUBLIC_ BOOL in_list(const char *s, const char *list, BOOL casesensitive)
519 while (next_token(&p,tok,LIST_SEP,sizeof(tok))) {
521 if (strcmp(tok,s) == 0)
524 if (strcasecmp_m(tok,s) == 0)
532 Set a string value, allocing the space for the string
534 static BOOL string_init(char **dest,const char *src)
538 (*dest) = strdup(src);
539 if ((*dest) == NULL) {
540 DEBUG(0,("Out of memory in string_init\n"));
549 _PUBLIC_ void string_free(char **s)
551 if (s) SAFE_FREE(*s);
555 Set a string value, deallocating any existing space, and allocing the space
558 _PUBLIC_ BOOL string_set(char **dest, const char *src)
561 return string_init(dest,src);
565 Substitute a string for a pattern in another string. Make sure there is
568 This routine looks for pattern in s and replaces it with
569 insert. It may do multiple replacements.
571 Any of " ; ' $ or ` in the insert string are replaced with _
572 if len==0 then the string cannot be extended. This is different from the old
573 use of len==0 which was for no length checks to be done.
576 _PUBLIC_ void string_sub(char *s,const char *pattern, const char *insert, size_t len)
581 if (!insert || !pattern || !*pattern || !s)
584 ls = (ssize_t)strlen(s);
585 lp = (ssize_t)strlen(pattern);
586 li = (ssize_t)strlen(insert);
589 len = ls + 1; /* len is number of *bytes* */
591 while (lp <= ls && (p = strstr(s,pattern))) {
592 if (ls + (li-lp) >= len) {
593 DEBUG(0,("ERROR: string overflow by %d in string_sub(%.50s, %d)\n",
594 (int)(ls + (li-lp) - len),
599 memmove(p+li,p+lp,strlen(p+lp)+1);
624 Similar to string_sub() but allows for any character to be substituted.
626 if len==0 then the string cannot be extended. This is different from the old
627 use of len==0 which was for no length checks to be done.
630 _PUBLIC_ void all_string_sub(char *s,const char *pattern,const char *insert, size_t len)
635 if (!insert || !pattern || !s)
638 ls = (ssize_t)strlen(s);
639 lp = (ssize_t)strlen(pattern);
640 li = (ssize_t)strlen(insert);
646 len = ls + 1; /* len is number of *bytes* */
648 while (lp <= ls && (p = strstr(s,pattern))) {
649 if (ls + (li-lp) >= len) {
650 DEBUG(0,("ERROR: string overflow by %d in all_string_sub(%.50s, %d)\n",
651 (int)(ls + (li-lp) - len),
656 memmove(p+li,p+lp,strlen(p+lp)+1);
658 memcpy(p, insert, li);
666 Strchr and strrchr_m are a bit complex on general multi-byte strings.
668 _PUBLIC_ char *strchr_m(const char *s, char c)
670 /* characters below 0x3F are guaranteed to not appear in
671 non-initial position in multi-byte charsets */
672 if ((c & 0xC0) == 0) {
678 codepoint_t c2 = next_codepoint(s, &size);
680 return discard_const(s);
689 * Multibyte-character version of strrchr
691 _PUBLIC_ char *strrchr_m(const char *s, char c)
695 /* characters below 0x3F are guaranteed to not appear in
696 non-initial position in multi-byte charsets */
697 if ((c & 0xC0) == 0) {
698 return strrchr(s, c);
703 codepoint_t c2 = next_codepoint(s, &size);
705 ret = discard_const(s);
714 return True if any (multi-byte) character is lower case
716 _PUBLIC_ BOOL strhaslower(const char *string)
723 s = next_codepoint(string, &c_size);
729 return True; /* that means it has lower case chars */
737 return True if any (multi-byte) character is upper case
739 _PUBLIC_ BOOL strhasupper(const char *string)
746 s = next_codepoint(string, &c_size);
752 return True; /* that means it has upper case chars */
760 Convert a string to lower case, allocated with talloc
762 _PUBLIC_ char *strlower_talloc(TALLOC_CTX *ctx, const char *src)
767 /* this takes advantage of the fact that upper/lower can't
768 change the length of a character by more than 1 byte */
769 dest = talloc_size(ctx, 2*(strlen(src))+1);
776 codepoint_t c = next_codepoint(src, &c_size);
781 c_size = push_codepoint(dest+size, c);
795 Convert a string to UPPER case, allocated with talloc
797 _PUBLIC_ char *strupper_talloc(TALLOC_CTX *ctx, const char *src)
806 /* this takes advantage of the fact that upper/lower can't
807 change the length of a character by more than 1 byte */
808 dest = talloc_size(ctx, 2*(strlen(src))+1);
815 codepoint_t c = next_codepoint(src, &c_size);
820 c_size = push_codepoint(dest+size, c);
834 Convert a string to lower case.
836 _PUBLIC_ void strlower_m(char *s)
840 /* this is quite a common operation, so we want it to be
841 fast. We optimise for the ascii case, knowing that all our
842 supported multi-byte character sets are ascii-compatible
843 (ie. they match for the first 128 chars) */
844 while (*s && !(((uint8_t)*s) & 0x80)) {
845 *s = tolower((uint8_t)*s);
855 size_t c_size, c_size2;
856 codepoint_t c = next_codepoint(s, &c_size);
857 c_size2 = push_codepoint(d, tolower_w(c));
858 if (c_size2 > c_size) {
859 DEBUG(0,("FATAL: codepoint 0x%x (0x%x) expanded from %d to %d bytes in strlower_m\n",
860 c, tolower_w(c), (int)c_size, (int)c_size2));
861 smb_panic("codepoint expansion in strlower_m\n");
870 Convert a string to UPPER case.
872 _PUBLIC_ void strupper_m(char *s)
876 /* this is quite a common operation, so we want it to be
877 fast. We optimise for the ascii case, knowing that all our
878 supported multi-byte character sets are ascii-compatible
879 (ie. they match for the first 128 chars) */
880 while (*s && !(((uint8_t)*s) & 0x80)) {
881 *s = toupper((uint8_t)*s);
891 size_t c_size, c_size2;
892 codepoint_t c = next_codepoint(s, &c_size);
893 c_size2 = push_codepoint(d, toupper_w(c));
894 if (c_size2 > c_size) {
895 DEBUG(0,("FATAL: codepoint 0x%x (0x%x) expanded from %d to %d bytes in strupper_m\n",
896 c, toupper_w(c), (int)c_size, (int)c_size2));
897 smb_panic("codepoint expansion in strupper_m\n");
906 Count the number of UCS2 characters in a string. Normally this will
907 be the same as the number of bytes in a string for single byte strings,
908 but will be different for multibyte.
910 _PUBLIC_ size_t strlen_m(const char *s)
918 while (*s && !(((uint8_t)*s) & 0x80)) {
929 codepoint_t c = next_codepoint(s, &c_size);
942 Work out the number of multibyte chars in a string, including the NULL
945 _PUBLIC_ size_t strlen_m_term(const char *s)
951 return strlen_m(s) + 1;
955 Unescape a URL encoded string, in place.
958 _PUBLIC_ void rfc1738_unescape(char *buf)
962 while ((p=strchr_m(p,'+')))
967 while (p && *p && (p=strchr_m(p,'%'))) {
971 if (c1 >= '0' && c1 <= '9')
973 else if (c1 >= 'A' && c1 <= 'F')
975 else if (c1 >= 'a' && c1 <= 'f')
977 else {p++; continue;}
979 if (c2 >= '0' && c2 <= '9')
981 else if (c2 >= 'A' && c2 <= 'F')
983 else if (c2 >= 'a' && c2 <= 'f')
985 else {p++; continue;}
989 memmove(p+1, p+3, strlen(p+3)+1);
995 * Decode a base64 string into a DATA_BLOB - simple and slow algorithm
997 _PUBLIC_ DATA_BLOB base64_decode_data_blob(TALLOC_CTX *mem_ctx, const char *s)
999 DATA_BLOB ret = data_blob_talloc(mem_ctx, s, strlen(s)+1);
1000 ret.length = ldb_base64_decode((char *)ret.data);
1005 * Decode a base64 string in-place - wrapper for the above
1007 _PUBLIC_ void base64_decode_inplace(char *s)
1009 ldb_base64_decode(s);
1013 * Encode a base64 string into a talloc()ed string caller to free.
1015 _PUBLIC_ char *base64_encode_data_blob(TALLOC_CTX *mem_ctx, DATA_BLOB data)
1017 return ldb_base64_encode(mem_ctx, (const char *)data.data, data.length);
1021 size_t valgrind_strlen(const char *s)
1024 for(count = 0; *s++; count++)
1032 format a string into length-prefixed dotted domain format, as used in NBT
1033 and in some ADS structures
1035 _PUBLIC_ const char *str_format_nbt_domain(TALLOC_CTX *mem_ctx, const char *s)
1040 return talloc_strdup(mem_ctx, "");
1042 ret = talloc_size(mem_ctx, strlen(s)+2);
1047 memcpy(ret+1, s, strlen(s)+1);
1050 for (i=0;ret[i];i++) {
1051 if (ret[i] == '.') {
1052 char *p = strchr(ret+i+1, '.');
1054 ret[i] = p-(ret+i+1);
1056 ret[i] = strlen(ret+i+1);
1065 * Add a string to an array of strings.
1067 * num should be a pointer to an integer that holds the current
1068 * number of elements in strings. It will be updated by this function.
1070 _PUBLIC_ BOOL add_string_to_array(TALLOC_CTX *mem_ctx,
1071 const char *str, const char ***strings, int *num)
1073 char *dup_str = talloc_strdup(mem_ctx, str);
1075 *strings = talloc_realloc(mem_ctx,
1077 const char *, ((*num)+1));
1079 if ((*strings == NULL) || (dup_str == NULL))
1082 (*strings)[*num] = dup_str;
1091 varient of strcmp() that handles NULL ptrs
1093 _PUBLIC_ int strcmp_safe(const char *s1, const char *s2)
1098 if (s1 == NULL || s2 == NULL) {
1101 return strcmp(s1, s2);
1106 return the number of bytes occupied by a buffer in ASCII format
1107 the result includes the null termination
1108 limited by 'n' bytes
1110 _PUBLIC_ size_t ascii_len_n(const char *src, size_t n)
1114 len = strnlen(src, n);
1124 Return a string representing a CIFS attribute for a file.
1126 _PUBLIC_ char *attrib_string(TALLOC_CTX *mem_ctx, uint32_t attrib)
1133 {'V', FILE_ATTRIBUTE_VOLUME},
1134 {'D', FILE_ATTRIBUTE_DIRECTORY},
1135 {'A', FILE_ATTRIBUTE_ARCHIVE},
1136 {'H', FILE_ATTRIBUTE_HIDDEN},
1137 {'S', FILE_ATTRIBUTE_SYSTEM},
1138 {'N', FILE_ATTRIBUTE_NORMAL},
1139 {'R', FILE_ATTRIBUTE_READONLY},
1140 {'d', FILE_ATTRIBUTE_DEVICE},
1141 {'t', FILE_ATTRIBUTE_TEMPORARY},
1142 {'s', FILE_ATTRIBUTE_SPARSE},
1143 {'r', FILE_ATTRIBUTE_REPARSE_POINT},
1144 {'c', FILE_ATTRIBUTE_COMPRESSED},
1145 {'o', FILE_ATTRIBUTE_OFFLINE},
1146 {'n', FILE_ATTRIBUTE_NONINDEXED},
1147 {'e', FILE_ATTRIBUTE_ENCRYPTED}
1151 ret = talloc_size(mem_ctx, ARRAY_SIZE(attr_strs)+1);
1156 for (len=i=0; i<ARRAY_SIZE(attr_strs); i++) {
1157 if (attrib & attr_strs[i].attr) {
1158 ret[len++] = attr_strs[i].c;
1168 Set a boolean variable from the text value stored in the passed string.
1169 Returns True in success, False if the passed string does not correctly
1170 represent a boolean.
1173 _PUBLIC_ BOOL set_boolean(const char *boolean_string, BOOL *boolean)
1175 if (strwicmp(boolean_string, "yes") == 0 ||
1176 strwicmp(boolean_string, "true") == 0 ||
1177 strwicmp(boolean_string, "on") == 0 ||
1178 strwicmp(boolean_string, "1") == 0) {
1181 } else if (strwicmp(boolean_string, "no") == 0 ||
1182 strwicmp(boolean_string, "false") == 0 ||
1183 strwicmp(boolean_string, "off") == 0 ||
1184 strwicmp(boolean_string, "0") == 0) {
1192 * Parse a string containing a boolean value.
1194 * val will be set to the read value.
1196 * @retval True if a boolean value was parsed, False otherwise.
1198 _PUBLIC_ BOOL conv_str_bool(const char * str, BOOL * val)
1203 if (str == NULL || *str == '\0') {
1207 lval = strtol(str, &end, 10 /* base */);
1208 if (end == NULL || *end != '\0' || end == str) {
1209 return set_boolean(str, val);
1212 *val = (lval) ? True : False;
1217 * Convert a size specification like 16K into an integral number of bytes.
1219 _PUBLIC_ BOOL conv_str_size(const char * str, uint64_t * val)
1222 unsigned long long lval;
1224 if (str == NULL || *str == '\0') {
1228 lval = strtoull(str, &end, 10 /* base */);
1229 if (end == NULL || end == str) {
1234 if (strwicmp(end, "K") == 0) {
1236 } else if (strwicmp(end, "M") == 0) {
1237 lval *= (1024ULL * 1024ULL);
1238 } else if (strwicmp(end, "G") == 0) {
1239 lval *= (1024ULL * 1024ULL * 1024ULL);
1240 } else if (strwicmp(end, "T") == 0) {
1241 lval *= (1024ULL * 1024ULL * 1024ULL * 1024ULL);
1242 } else if (strwicmp(end, "P") == 0) {
1243 lval *= (1024ULL * 1024ULL * 1024ULL * 1024ULL * 1024ULL);
1249 *val = (uint64_t)lval;
1254 * Parse a uint64_t value from a string
1256 * val will be set to the value read.
1258 * @retval True if parsing was successful, False otherwise
1260 _PUBLIC_ BOOL conv_str_u64(const char * str, uint64_t * val)
1263 unsigned long long lval;
1265 if (str == NULL || *str == '\0') {
1269 lval = strtoull(str, &end, 10 /* base */);
1270 if (end == NULL || *end != '\0' || end == str) {
1274 *val = (uint64_t)lval;